-3

I have HP DL380 G7, with 8x300GB disks. The disks are not HP genuine, they are OEM.

I configured RAID 0 for all disks, as i need speed and i have backup plans.

within 1 month: 1 Disk threw SMART error, and 1 disk completely failed (aka click of death). another 1 month, another disk threw SMART error.

Now the question:

Based on your experiences, If i gonna purchase new genuine hard drives from HP store, will they failed within 3 months because i setup 6 to 8 disk as single RAID 0 volume ? (failed means either click-of-death or SMART error).

Thanks.

Hongkie
  • 3
  • 2
  • 2
    As we make very clear when you sign up for serverfault.com is for sysadmins working in a professional environment. Nothing you have done here fits the description of professional - consider adopting a degree of professionalism in your work, or look towards a career where corner-cutting doesn't have such a cataclysmic series of outcomes. – Chopper3 Feb 16 '18 at 11:25
  • 1
    @Hongkie: We are not here to only tell you what you want to hear. This is a site for professional system administration and buying used crap from ebay doesn't fit this description. – Sven Feb 16 '18 at 13:53
  • i am a professioinal admin, with a limited budget. small company. Are you telling me only Fortune 500 system admin can post here ? – Hongkie Feb 16 '18 at 14:00
  • 1
    No, it doesn't mean this. You can have a small budget and still work professionally, but there is a point where the professional approach is "with this amount of money, we can't do this project. Either give me more or don't do it at all". – Sven Feb 16 '18 at 14:08
  • no sven, in real world, Client always has small budget and company looks for business by bidding. Client cannot afford Gen9, therefore they look for G6 or G7 class server and solutions... for a fast hard drive that enough not for long term, but for short term... you cannot generalize everything – Hongkie Feb 16 '18 at 14:17
  • 1
    I'm sorry Hongkie but you're wrong - what you've done isn't professional, it's about standards of professional behaviour and you're simply trying to justify bad behaviour - we all have financial limitation (even big companies), it's about how you approach your work and responsibilities. If you're going to hack something together have the guts to admit that that is what it is and that what you've done is unprofessional, anything else is dishonest or deluded. – Chopper3 Feb 16 '18 at 14:32
  • wrong chopper. US gaves business to chinese is because the labour are cheaps. Hard drives that comes from china also cheaps. And many people are earn living for that. Buying from China / ebay is not bad behaviour, nor needs justification – Hongkie Feb 16 '18 at 15:00
  • 1
    We're not going to agree, perhaps this place isn't for you. – Chopper3 Feb 16 '18 at 15:07
  • Make this question go away. – ewwhite Feb 17 '18 at 06:46

1 Answers1

2

The RAID0 has nothing to do with the failures and if you operate good new drives within specifications (e.g. not in extreme environmental conditions and mechanically stable), they should last for years. Also, if they fail within a short time, you usually are protected by some kind of warranty.

Beside that, consider buying SSDs if you need speed.

Sven
  • 97,248
  • 13
  • 177
  • 225
  • Thanks for SSD advise. It's out of budget. Would you mind explain: What is extreme environmental condition and mechanically stable ? The server stays in room temperature. Ambient temperature around 27 - 31 degree celcius. Server stays in clean room. – Hongkie Feb 16 '18 at 13:26
  • Look up the disks datasheet. It will list valid operating conditions, stay within them. By mechanical stable I mean that the server is sitting stable in an unmoving rack, without external vibrations etc. – Sven Feb 16 '18 at 13:29
  • The original 300GB has HP Part No: 652564-B21 https://h20195.www2.hpe.com/v2/getpdf.aspx/c04111744.pdf?ver=50 based on the link above, the only "common sense" operating condition is the temperature. It is 10 degree to 35 degree celsius. now the problem is... these OEM drives that i have seems to "blind" the sea of sensors... (fyi: HP sea of sensor)... temperature always 50 degree, but i doubt it's 50 degree... do you think the temperature kills my 3 drives earlier ? – Hongkie Feb 16 '18 at 13:57
  • 1
    They specify "system inlet air temperature", meaning the ambient temperature of the area outside the air intake of the server. The sensors in a server usually measure internal temperatures and thus values are higher, but the idea is that the servers and drives are designed to not overheat if the incoming "cold" air is within the specified temperature range (and the airflow within the server is not restricted by dirt, broken fans or user modifications). – Sven Feb 16 '18 at 14:06
  • ok i see the points. Thanks for clarifying RAID0 has nothing to do with failures. – Hongkie Feb 16 '18 at 14:23
  • @Hongkie Non-HP and consumer disks aren't intended for use in these servers. If you are using inappropriate disks, the temperature sensors may not work. This is a bad idea. – ewwhite Feb 17 '18 at 06:47
  • @ewwhite, 2 things: 1. I assumed you mean consumer disks as SATA ? no, all disks i am talking about are SAS. 2... Also HP drives are outsourced to many vendor. eg. seagate savvio... i dont wanna say seagate is ugly so i simply use word: OEM – Hongkie Feb 17 '18 at 07:05